NVIDIA designs accelerated computing platforms and sells them as silicon, boards, modules, systems, and software. Revenue is led by Data Center, with Gaming as the second pillar, plus smaller platforms in Professional Visualization and Automotive. In Q4 fiscal 2026 (ended January 25, 2026), NVIDIA reported $68.1 billion of revenue, including $62.3 billion from Data Center, $3.7 billion from Gaming, $1.3 billion from Professional Visualization, and $604 million from Automotive. Full-year fiscal 2026 revenue reached $215.9 billion, with Data Center at $193.7 billion, Gaming at $16.0 billion, Professional Visualization at $3.2 billion, and Automotive at $2.3 billion.
Kernaktivitäten
- Data Center compute platforms
NVIDIA sells GPUs and integrated platforms for AI training and inference, analytics, graphics, and scientific computing. Products ship as PCIe cards, modules, and complete systems through cloud providers, OEMs, and server makers. In Q4 fiscal 2026, Data Center compute revenue was $51.3 billion, reflecting the ramp of Blackwell-based platforms.
- Networking for AI clusters
NVIDIA sells InfiniBand and Ethernet networking for AI clusters, spanning adapters, switches, cables, DPUs, and fabric software. Spectrum-X is positioned as its Ethernet platform for AI data centers. In Q4 fiscal 2026, Data Center networking revenue was $11.0 billion.
- Rack-scale systems
NVIDIA packages chips into rack-scale architectures using NVLink and system-level reference designs that scale across many GPUs, aimed at large training and inference clusters.
- Software licensing and platform services
CUDA provides the programming model and libraries that underpin NVIDIA’s AI and HPC stack. Enterprise software is monetized through licensed offerings such as NVIDIA AI Enterprise and Omniverse Enterprise, sold with support and deployment tooling.
- Graphics and edge markets
Gaming is driven by GeForce GPUs and services such as GeForce NOW. Professional Visualization centers on RTX workstation GPUs and related software. Automotive revenue is tied to DRIVE hardware and software sold into long program cycles with carmakers and suppliers.
Marktposition
NVIDIA’s moat comes from controlling compute, networking, systems, and software as one stack, which supports platform-level optimization and faster product refresh cycles. CUDA and its library ecosystem anchor developer adoption and raise switching costs for production AI workloads. Data Center results show the scale of that stack, with Q4 fiscal 2026 Data Center revenue split between compute ($51.3B) and networking ($11.0B).
Roadmap visibility extends beyond Blackwell. NVIDIA unveiled the Rubin platform and signaled early cloud deployments, with partner rollouts starting in the second half of 2026. In its Q1 fiscal 2027 outlook, NVIDIA stated it assumes no Data Center compute revenue from China, which frames geopolitics and export rules as a key constraint on incremental growth.